This function is used internally by epinowcast to replace
default model priors with users specified ones (restricted to
normal priors with specified mean and standard deviations). A common
use would be extracting the posterior from a previous epinowcast()
run (using summary(nowcast, type = fit)) and using this a prior.
Arguments
- priors
A
data.framewith the following variables:variable,mean,sddescribing normal priors. Priors in the appropriate format are returned byenw_reference()as well as by other similar model specification functions.- custom_priors
A
data.framewith the following variables:variable,mean,sddescribing normal priors. Priors in the appropriate format are returned byenw_reference()as well as by other similar model specification functions. Priors in this data.frame replace the default priors. Note that currently vectorised prior names (i.e those of the formvariable[n]will be treated asvariable).
See also
Functions used to help convert models into the format required for stan
enw_formula_as_data_list(),
enw_get_cache(),
enw_model(),
enw_pathfinder(),
enw_priors_as_data_list(),
enw_sample(),
enw_set_cache(),
enw_stan_to_r(),
enw_unset_cache(),
remove_profiling(),
write_stan_files_no_profile()
Examples
# Update priors from a data.frame
priors <- data.frame(variable = c("x", "y"), mean = c(1, 2), sd = c(1, 2))
custom_priors <- data.frame(variable = "x[1]", mean = 10, sd = 2)
enw_replace_priors(priors, custom_priors)
#> variable mean sd
#> <char> <num> <num>
#> 1: y 2 2
#> 2: x 10 2
# Update priors from a previous model fit
default_priors <- enw_reference(
distribution = "lognormal",
data = enw_example("preprocessed"),
)$priors
print(default_priors)
#> variable
#> <char>
#> 1: refp_mean_int
#> 2: refp_sd_int
#> 3: refp_mean_beta_sd
#> 4: refp_sd_beta_sd
#> 5: refnp_int
#> 6: refnp_beta_sd
#> description
#> <char>
#> 1: Log mean intercept for parametric reference date delay
#> 2: Log standard deviation for the parametric reference date delay
#> 3: Standard deviation of scaled pooled parametric mean effects
#> 4: Standard deviation of scaled pooled parametric sd effects
#> 5: Intercept for non-parametric reference date delay
#> 6: Standard deviation of scaled pooled non-parametric effects
#> distribution mean sd
#> <char> <num> <num>
#> 1: Normal 1.0 1
#> 2: Zero truncated normal 0.5 1
#> 3: Zero truncated normal 0.0 1
#> 4: Zero truncated normal 0.0 1
#> 5: Normal 0.0 1
#> 6: Zero truncated normal 0.0 1
fit_priors <- summary(
enw_example("nowcast"), type = "fit",
variables = c("refp_mean_int", "refp_sd_int", "sqrt_phi")
)
fit_priors
#> variable mean median sd mad q5 q20
#> <char> <num> <num> <num> <num> <num> <num>
#> 1: refp_mean_int[1] 2.9050494 2.8662550 0.6298971 0.60792530 1.906698 2.374370
#> 2: refp_sd_int[1] 3.7061514 3.6991850 0.3012037 0.30680183 3.237901 3.458442
#> 3: sqrt_phi[1] 0.3465375 0.3448305 0.0314694 0.03032955 0.294676 0.321174
#> q80 q95 rhat ess_bulk ess_tail
#> <num> <num> <num> <num> <num>
#> 1: 3.4255460 4.0286135 0.9998312 856.6469 768.7502
#> 2: 3.9609120 4.2112490 0.9990012 900.2061 844.7802
#> 3: 0.3735868 0.4001852 1.0005785 873.5007 771.9195
enw_replace_priors(default_priors, fit_priors)
#> variable
#> <char>
#> 1: refp_mean_beta_sd
#> 2: refp_sd_beta_sd
#> 3: refnp_int
#> 4: refnp_beta_sd
#> 5: refp_mean_int
#> 6: refp_sd_int
#> 7: sqrt_phi
#> description
#> <char>
#> 1: Standard deviation of scaled pooled parametric mean effects
#> 2: Standard deviation of scaled pooled parametric sd effects
#> 3: Intercept for non-parametric reference date delay
#> 4: Standard deviation of scaled pooled non-parametric effects
#> 5: <NA>
#> 6: <NA>
#> 7: <NA>
#> distribution mean sd
#> <char> <num> <num>
#> 1: Zero truncated normal 0.0000000 1.0000000
#> 2: Zero truncated normal 0.0000000 1.0000000
#> 3: Normal 0.0000000 1.0000000
#> 4: Zero truncated normal 0.0000000 1.0000000
#> 5: <NA> 2.9050494 0.6298971
#> 6: <NA> 3.7061514 0.3012037
#> 7: <NA> 0.3465375 0.0314694